Administrator/Researcher: Angela J. Cone
Costa Tsirigakis - Founder J2 Y DNA project & admin/researcher from 2006 - mid 2008
See bottom of page for Project Admin. information

Analysis Phase 2
  Marker distributions in J2 sub-clades
   First 12 markers
   • Second results panel
   • Third results panel
Analysis Phase 3
  J2 Cluster analysis
   Marker panel comparison
   • Cluster analysis
   • J2b-Delta hypotheses

J2 cluster analysis
(The usefulness of each individual marker panel is examined, and clusters within J2a and J2b are defined)

The J2 Y-DNA project phase three analysis was performed for three reasons:

1) To closely examine the differences in sub-clade predictability between the first three FTDNA marker panels.
2) To expand upon and verify the findings of the phase 1& 2 analyses, and present the results in a "user friendly" manner.
3) To investigate some interesting/unusual observations in the Network diagrams with regard to J2 M12+ (J2b - Sengupta et al. 2006; J2a - Regueiro et al. 2006; J2e - YCC)

These results will be presented in two main parts:

A: FTDNA Marker panel comparison

B: J2 Y-DNA Cluster analysis

 

A: FTDNA Marker panel comparison

The J2 Y-DNA marker panel comparison was conducted in response to enquiries from project members, asking why their position in the Network diagrams (relative to other project members) seemed to alter between the 12, 25 and 37 marker Network diagrams.
These analyses were conducted using the 102 haplotypes in the project that had results from all markers in each of the first three marker panels (mid December). Network diagram analysis "A" included markers 1 -12; network diagram analysis "B" included markers 13-21 (DYS 464 was excluded from the analysis due to its status as a multicopy marker); Network diagram analysis "C" included markers 26-27.
These separate network diagram analyses for each of the first three FTDNA marker panels illustrate how the different marker panels differ in their usefulness for distinguishing subclades.

Two additional network diagram analyses were conducted which verify the summarized results in phase 2 of the analysis. Network Diagram analysis "D" was performed using the markers that were determined as the least useful for distinguishing between subclades, and the final diagram (below) was conducted using the markers that were determined as being of the greatest value in distinguishing between different sub-clades.

One aspect of the analysis below that should be made very clear is:
We would expect haplotypes from the separate subclades to cluster together most closely in this analysis, since the markers used in the analysis were selected by determining which markers were the most statistically significant in distinguishing between the clades.
Therefore, what this diagram does is merely confirm the validity of the phase two analysis results
(if on the other hand the sub-clades had not clustered together, it would have challenged the overall validity of the phase two analysis results).

When examining these separate network diagram analyses, - keep in mind that the exact same set of haplotypes were used in each of them. Each diagram includes the exact same haplotypes that were used in the 37 marker diagram analysis. The only difference, is the subset of the DYS markers that were used.

It can be seen that in most of the analyses, nodes of a particular colour clump together. The different node colours each represent a different J2 subclade (as determined by SNP testing). It can be seen that in some marker panels the clumping is more distinct, and in other marker panels it is less distinct. It can also be seen that there is no discrete clumping in the network diagram that is derived from the least significant markers, and the clumping is most distinct in the network diagram that is derived from the most significant markers. It can also be seen that of the three marker panels the clumping is least distinct for the first marker panel and the most distinct for the third marker panel.

Of special note- It can be seen that the markers in the first marker panel are most useful for distinguishing J2 M67+/M92+ (J2a2a Karafet et al. 2008)., and the markers in the second marker panel are most useful for distinguishing J2 M205+ (J2b1 Karafet et al. 2008). Results to date suggest it is quite difficult to distinguish M12+/M205+ from M12+/M205- in the absence of the markers in the second marker panel (and most specifically markers DYS 454 and DYS 437).

One overall conclusion that can be drawn from this analysis is that (in J2), the first 12 markers are of quite limited use in predicting subclade - to get maximum information about a haplotype, at least 25 markers are needed, preferably more.

B: J2 Y-DNA Cluster analysis

The J2 Y-DNA cluster analysis was conducted to expand upon the results of the above marker panel analysis, and also to present the results of the first two analysis phases into a form that has greater practical application (and hopefully in a form that is more easily understood by the average person).
An additional impetus for conducting this analysis was to verify an apparent correlation between a sub-grouping of J2~M12 haplotypes and geographical origin.

This analysis attempts to replicate (in a modified manner) the system of clustering that was given to E3b (Now E1b1b1) in published scientific studies (ie. Cruciani et al. 2004; Cruciani et al 2006)
The same system of assigning Mathematical* Greek letters to the different clusters will be used. To avoid confusion with the sub-clade nomenclature, Greek letters that closely resemble Roman letters are omitted.

In defining the clusters, multiple methods of clustering analyses were performed, and the results of the different methods were compared. Attention was paid to whether the clustering patterns were consistent between the different analyses. This was a time consuming process (and the primary reason for the delay in presenting these results). This was done to ensure that the defining of clusters was as scientifically objective as possible. The output of these different clustering methods will be presented at a later date.

Before the results of the clustering analysis are presented, there is one fact that needs to be made clear:

**A cluster is not the same as a clade**

A cluster = a group of similar haplotypes
A clade = branch of the J2 "family tree"

As an analogy, - It's like the difference between constructing a family tree by the physical appearance of each individual v's a family tree constructed from genealogical records.
On a relative scale, - STR's mutate quite rapidly, - which is precisely why they are useful for assisting us in determining recent relationships (ie. in genealogical research). However, this fast mutation rate makes them progressively less scientifically accurate/objective for assessing the relationships between haplotypes the further back in time we go. All that analyses with haplotypes can do is represent haplotype similarity, rather than the phylogenetic relationships between the haplotypes.

Often haplotype clustering will identify groups of haplotypes that are likely to be phylogenetically related, but the relationships between the clusters are often depicted in a way that is contrary to known phylogenetic relationships. This is not due to deficiencies in the algorithms employed in analyzing the STR haplotypes, - this is due to the inherent nature of STR's in being less accurate in assessing relationships between lineages the further back in time we go. The deeper we go in ancestry (ie. anthropological time scale) the less useful STR's are, and the more useful SNP's are; the shallower we go in ancestry (ie. genealogical time scale) the more useful STR's are, and the less useful SNP's are.

Therefore a cluster ≠ clade

Within a cluster, some of the haplotypes will be closely related phylogenetically, but others may not. Some clusters may consist of haplotypes within a single haplogroup clade, and some may contain haplotypes from more than one haplogroup clade. In a way, what the clustering does is help us create hypotheses which can then be tested (eg. by subclade testing with currently known SNP's, and SNP's that have yet to be discovered in the future).

37 marker diagram (analysed December 2006) showing defined clusters

 

 

Within J2b/M12+ (Karafet et al. 2008) there are three main clusters defined:


J2b-α (J2b-Alpha) appears to be equivalent to J2b M102+/M205- J2b(xJ2b1) Karafet et al. 2008 ).

•• J2b-δ (J2b-Delta) is a sub-cluster within J2b-Beta.
An interesting observation is that 7 of the 9 members in cluster J2-δ have a paternal line geographical origin from the British isles (one other has a geographical origin of Spain, and the other is unknown). Much of the extra analysis was done in the effort to verify whether this observation was a "biologically meaningful" or a spurious correlation. The overall conclusion at present is that more results are needed before we can fully determine whether this means anything. This apparent correlation is further discussed here

••• J2-ζ (J2b-Zita) is a sub-cluster within J2b-Delta.
This cluster is distinguished by a value of 10 at DYS 385a. One of the two members in this cluster has a paternal line origin from Great Britain, and the paternal line origin of the other member is unknown. There are three other project members who have a value of 10 at DYS 385a, - and it is not known whether these members would also fall into the same cluster if their haplotype had 37 markers. (these other members have origins outside of the British Isles, so if they are in J2b-Zita, and therefore also in J2b-Delta, it might challenge the apparent correlation between J2b-Delta and the British Isles).
One member in J2b-Zita has unusual results* in the 38-67 marker panel, so it would be valuable to know whether others in J2b-Zita also have these unusual results or not.


J2b-β (J2b-Beta) appears to be equivalent to J2b M102/M205+ J2b1 - Karafet et al. 2008 ).
These are preliminary findings only, - more extended M205+ haplotypes need to be known before we can make any scientifically firm/objective conclusions.
Within the project there are 6 haplotypes in J2b-alpha, four of which are SNP confirmed as M205+ (the remaining two are not yet SNP tested). The markers that most reliably predict a haplotype is J2b-β are found in the second FTDNA results panel, - so at least 25 markers are needed to reliably identify this cluster.
As yet there is no data on M205+ in published DNA studies, so its distribution and the potential age of the cluster has not yet been calculated in Academic research.
The geographical origins of the 6 members in J2b-alpha are: Serbia, Mexico, Greece (2 haplotypes), Germany, and England. Of note is the fact that there are genetic distances of 14-23 at 37 markers, which implies that the clade may have been founded at least 4000 years ago (assuming an average mutation rate of .002, and an average generation time of 27 years).
J2b-γ (J2b-Gamma) also appears to be J2b M102+/M205- J2b(xJ2b1) Karafet et al. 2008 ).
It remains to be determined whether there is any phylogenetic distinction between the members who fall within J2b-Alpha and J2b-Gamma.

Within J2a/M410+ (Karafet et al. 2008) there are three main clusters defined:

Within the project, the J2a members that have been SNP tested have been found to be either M67+ or negative for all of the SNP markers that they have been tested for that are "down stream" of M410. Below we refer to haplotypes as being either M67+ or M67-. This is done for simplicity and in fact most of the members stated to be M67- have also tested negative for M47, M68, M137, M158, & M339.

J2a-α (J2a-Alpha) broadly contains both M67+ and M67- haplotypes. Those that are M67- are often difficult to distinguish from those that are M67+. Most of these haplotypes are likely to be DYS 413 ≤ 18, however some may be DYS 413 ≥ 18. There are two confirmed DYS 413 ≥ 18 haplotypes that are outside of J2a-Alpha that branch off the J2a-Alpha cluster (the results of one of these was received after the 37 marker analysis was completed).

•• J2a-δ (J2a-Delta) is a sub-cluster within J2a- Alpha.
This cluster appears to contain haplotypes that are M67+.

••J2a-σ (J2a-Sigma) is a sub-cluster within J2a-Delta.
This cluster contains very closely related haplotypes that are M67+. It is likely that most members who are in this cluster shared an ancestor within the last 2000 years. Many of the members in this cluster have paternal line origins from Hungary, Poland, Ukraine, or Lithuania.

••J2a-ω (J2a-Omega) is a sub-cluster within J2a-Delta.
This cluster contains haplotypes that are M67+

••J2a-π (J2a-Pi) is a sub-cluster within J2a-Alpha.
This cluster contains haplotypes that are M67+, M92+

•••J2a-φ (J2a-Phi) is a sub-cluster within J2a-Pi
This cluster contains very closely related haplotypes that are M67+, M92+. It is likely that most members who are in this cluster shared an ancestor within the last 2000 years. Many of the members in this cluster have paternal lines from Belarus, Lithuania, and also Ukraine.

••J2a-ξ (J2a-Xi) is a sub-cluster within J2a-Alpha.
At present insufficient haplotypes within this cluster have been SNP tested. Members in this cluster have paternal line origins in Greece, and the British isles

•••J2a-ψ (J2a-Psi) is a subcluster within J2a-Phi.
At present insufficient haplotypes within this cluster have been SNP tested. Members in this cluster have paternal line origins in the British Isles.

••J2a-μ (J2a-Mu) is a subcluster within J2a-Alpha.
At present insufficient haplotypes within this cluster have been SNP tested. Members in this cluster have paternal line origins in the British isles.
** This cluster now appears to M67+ (added: 1 Dec 2007)

••J2a-η (J2a-Eta) is a subcluster within J2a-Alpha.
At present insufficient haplotypes within this cluster have been SNP tested.
** We now suspect that this cluster is M92+ (added: 1 Dec 2007)

••J2a-ζ (J2a-Zita) is a subcluster within J2a-Alpha.
This cluster appears to contain haplotypes that are M67-
Members in this cluster have paternal line origins in Italy and Switzerland.
** We now suspect that this cluster may fall outside of J2a-Alpha (added: 1 Dec 2007)

J2a-β (J2a-Βeta) contains M67- haplotypes. Most (if not all) are likely to have DYS 413 ≤ 18.

••J2a-λ (J2a-Lamda) is a subcluster within J2a-Beta.
This cluster appears to contain haplotypes that are M67-
Members in this cluster have paternal line origins in Bohemia (1), France (1), Germany (5), British Isles (4), Italy (5), Malta (1), the Netherlands (1), and Switzerland (1).

••J2a-θ (J2a-Theta) is a subcluster within J2a-Beta.
This cluster appears to contain haplotypes that are M67-
Some of the members in this cluster are likely to have shared an ancestor within the last 2000 years.
Members in this cluster have paternal line origins in the Ukraine (3), Italy (2), and Russia.

J2a-γ (J2a-Gamma) contains M67+ haplotypes that branch closely to J2a-β. These M67+ haplotypes are distinguished from the M67+ haplotypes in J2a-α by DYS 459 values of 8,9.

Modal Values

Modal values are given for clusters that have at least 4 haplotypes.

J2b

(M12/M102+)

3
9
3
3
9
0
1
9
3
9
1
3
8
5
a
3
8
5
b
4
2
6
3
8
8
4
3
9
3
8
9
-
1
3
9
2
3
8
9
-
2
4
5
8
4
5
9
a
4
5
9
b
4
5
5
4
5
4
4
4
7
4
3
7
4
4
8
4
4
9
        4
6
0
G
A
T
A
H
4
Y
C
A
I
I
A
Y
C
A
I
I
B
4
5
6
6
0
7
5
7
6
5
7
0
C
D
Y
a
C
D
Y
b
4
4
2
4
3
8
J2b-α   24 15 10 14 17     12 12   28 16 8 9   11 28 16 19 29         11   19 20 13 14 16 17     11 9
  • J2b-δ   24 15 10 13 17     12 12   28 16 8 9   11 28 16 19 29         11   19 20 13 14 16 17     11 9
J2b-β   24 15 10 16 18     12 12   28 17 8 9   12 25 14 19 29         11   19 20 14 14 18 18     11 9

 

J2a

(M410+)

3
9
3
3
9
0
1
9
3
9
1
3
8
5
a
3
8
5
b
4
2
6
3
8
8
4
3
9
3
8
9
-
1
3
9
2
3
8
9
-
2
4
5
8
4
5
9
a
4
5
9
b
4
5
5
4
5
4
4
4
7
4
3
7
4
4
8
4
4
9
        4
6
0
G
A
T
A
H
4
Y
C
A
I
I
A
Y
C
A
I
I
B
4
5
6
6
0
7
5
7
6
5
7
0
C
D
Y
a
C
D
Y
b
4
4
2
4
3
8
J2a   23 14 10 13 16     11 13   29 15 9 9   11 26 15 20 29         10   19 22 15 14 16 17     12 9
  • J2a-α   23 14 10 13 17     11 13   31 18 9 9   11 26 15 20 29         10   19 22 15 13 19 17     12 9
     • J2a-δ   23 14 10 13 17     11 13   31 18 9 9   11 26 15 20 29         10   19 22 15 13 19 16     12 9
       • J2a-σ   23 14 10 13 17     11 13   30 18 9 9   11 26 15 20 29         10   19 22 15 13 19 17     12 9
     • J2a-π   22 14 10 14 15     12 14   31 15 9 9   11 26 15 20 29         10   19 22 15 12 15 17     12 9
       • J2a-φ   22 14 10 14 15     12 14   31 15 9 9   11 26 15 20 29         10   19 22 15 12 15 17     12 9
     • J2a-ξ   23 14 10 13 17     11 13   29 14 9 9   11 24 15 22 29         10   19 22 15 15 17 14     11 9
  • J2a-β   23 15
9
14 16     12 13   29 15 8 9   11 26 14 21 32         10   19 22 16 14 16 18     12 9
     • J2a-λ   23 15
9
13 16     11 13   29 14 8 9   11 26 14 21 32         10   19 22 16 14 16 18     12 9
     • J2a-θ   23 15 10 14 17     12 13   29 15 8 9   11 24 15 21 32         10   19 23 16 14 18 17     12 9

The Marker values are omitted for the markers that were determined to be of little significance in distinguishing between clades in the first phase analysis. The marker values for the markers determined to be of the greatest significance are in black text, and those of medium significance in white text.

In some instances there are clear modal differences for some medium significance markers between the clusters. The significance of these markers may have been underestimated in the phase two analysis, as they may (for instance) signify differences between groups that have no defining SNP marker (and the phase 2 analysis was dependant on analysing differences in marker values between objectively defined phylogenetic groupings (ie. sub-clades identified by SNP markers)). The project is aware that some markers may be of greater true significance in identifying biological groupings that the current phase 2 analysis results suggest. The phase 2 analysis will be re-analysed at an appropriate time in the future.

J2b-Delta & the British Isles

It's important to remember that this is science in progress, - sometimes when scientific hypotheses are researched the results prove the hypothesis, sometimes they disprove the hypothesis, and sometimes the initial results are inconclusive (thus indicating more research is needed). At this stage, we believe there is a 50:50 chance that the hypotheses presented below will be proved v's disproved.
Testing the below hypotheses will be an ongoing process, - and whatever the outcome, the process of testing these hypotheses will be a "behind the scenes" demonstration of real science in action. Also bear in mind that in the real scientific world, there is probably a bias towards publishing studies where the hypotheses are proven (but not because the scientists want to suppress negative results, but more because the results might be inconclusive, and the scientists might want to "go back to the drawing board", and think of a better way of testing their hypothesis - afterall, if the scientist tried to publish their incomplete results in a journal the peer reviewers would probably tell the journal that better results are needed before the study can be published).

The apparent correlation between J2b-Delta and geographic origin from the British Isles cannot as yet be 100% confirmed scientifically (nor can it be disproved).
We currently lack (for example) SNP marker evidence which might tell us that J2b-Delta represents a separate biological lineage from the other haplotypes within J2b-Alpha. Further J2b haplotypes with at least 37 markers will also help evaluate whether the apparent association with the British origins remains (or whether it was merely co-incidental).

The main distinguishing marker value for members within J2 is a value of 13 at DYS 385a. We cannot yet be sure whether the members within this group share this value largely as a result of descent from a common ancestor, or by convergence. If it is the former, then this grouping may represent a clade within J2b, and if it is the latter then this cluster just represents a spurious clustering pattern that is not underpinned by a true phylogenetic relationship.

Therefore, regard this apparent correlation as a hypothesis (not a fact) that requires additional scientific evidence before it can be proven (or disproved). Whether or not J2b-Zita corresponds to a discrete biological lineage that falls within J2b-Delta is also as yet unconfirmed. It might cluster with J2b-Delta due to its value of 10 at DYS 385a, but be otherwise not closely related biologically to the rest of J2a-Delta.

Below we formally outline the observation made, the deductions made from the observations, and the hypotheses that have been made to scientifically test the deductions made from the observations.

Observation: Almost all members within a J2b cluster in the 37 marker Network diagrams, stated a paternal origin from somewhere in the British Isles. This observation suggested that there might be a correlation between the J2b-Delta cluster and an origin in the British Isles. Two hypotheses have been constructed to test whether this apparent correlation is biological fact, or whether it was merely a spurious observation.

Hypothesis 1 : "J2b-Delta cluster represents a true biological lineage (ie. clade) within J2b"

This hypothesis could be proven or disproved on the basis of the results of testing SNP markers that may help define true biological clades (ie. "phylogenetic clades", or "true branches of the J2b family tree"). This could be done either with existing SNP markers, or as yet unidentified SNP markers that may be found in the future.

It is unfortunate that one of the main testing companies does not test for M241, - as knowing which J2b haplotypes are M241+ and which are M241- would be of extreme value to the project.[This DNA company now does test for M241+]
M241 is a marker that is used in numerous peer reviewed scientific publications, so its non inclusion in the YCC tree is puzzling.
One such scientific study was Pericic et al. 2005, which found that approximately half of J2b haplotypes sampled were M241+. Albanians were found to all be M241+, and Serbians were largely M241-. Similarily, Cinnioglu et al. 2004 and Regueiro et al. 2006 also found that about half of the M12/M102+ haplotypes tested were M241+. In contrast, Sengupta et al. 2006 found all J2b haplotypes in India to be M241+.

These studies did not test for M205, so some of these M241- haplotypes may have been M205+. None of these studies tested for markers that are suggestive of M205 so it is not possible to predict what proportion of the M241- haplotypes may have been M205--. It is unfortunate that studies to date have not tested for M205, as it would be of great interest to know what percentage of J2b haplotypes are both M205- and M241- (and accordingly interesting to know the percentages of M205+ in different regions).

It may be (for instance) that J2b-Delta is largely M241+ and the J2b-Alpha haplotypes outside of J2b-Delta are largely M241- (or vice versa). On the other hand, it may be found that M241- and M241+ are found evenly in the two groupings (which would disprove the hypothesis that J2-Delta is a discrete biological lineage).
More J2b haplotypes tested for M241 are needed to help prove or disprove the hypothesis that the J2b-Delta cluster represents a true biological lineage (ie. a M205- subclade within J2b).

 

Hypothesis 2 : "J2b-Delta is more closely associated with the British Isles than any other geographical location "

Proving (or disproving) hypothesis two will be less straightforward than proving (or disproving) hypothesis one. Hypothesis one can be tested using scientifically objective criteria, however, proving or disproving hypothesis two will be comparatively subjective, - so therefore extreme care needs to be taken in evaluating hypothesis two (it may be impossible to prove or disprove 100% beyond a doubt). If hypothesis one is invalid, then it would logically follow that hypothesis 2 is also invalid. Therefore to evaluate the apparent correlation between J2b-Delta cluster and ancestry from the British Isles,our first objective is to verify hypothesis two, whilst also determining the most objective methods for evaluating hypothesis 2.

One method of evaluating this hypothesis will be to continue to compare the %'s of members origins in J2b-Delta compared to others in J2b-Alpha, and also compared to those in the other J2b clusters, and in J2 overall. These figures should also be compared to the relative proportions of testees globally from different ancestral origins (eg. the degree to which the overall databases are biased towards British ancestry, compared to origins elsewhere, and how this bias may affect the pattern of results seen).
More J2b haplotypes tested for 37 markers are needed to help prove or disprove the hypothesis that the J2b-Delta cluster is more closely associated with the British Isles than any other geographical location.

Possible Implications:

If J2b-Delta is a separate discrete lineage within J2b, this could have implications regarding when the lineage arrived in the British Isles. At present the two main hypotheses regarding the arrival of J2 in the British Isles that are commonly touted are

1) Neolithic farmers

2) Roman soldiers

The first hypothesis is derived from scientific & archeological evidence that closely associates the early initial spread of agriculture with the distribution and spread of Haplogroup J2, and more specifically with J2a. This logically leads to the conclusion that if the association was between agriculture and J2a specifically, then it does not necessarily follow that J2b was introduced into populations by neolithic farmers.

It is a fact that the DNA databases are biased towards ancestral lines that originate from the British isles, so the apparent correlation could in part be an artificial result derived from the geographical bias in the databases,.. but on the other hand, there are no other clusters within J2 that have such a strong apparent correlation with the British Isles.

If it is a true clade, and is most closely associated with the British Isles, we would need to assess the likely age of the cluster (from haplotype variance), which might give us an indication of how long it might have been in the British Isles. If it is unlikely that its presence in the British Isles is the result of Neolithic farmers, nor Roman soldiers, then alternative hypotheses for its arrival in the British Isles would need to be evaluated.

April 2007 Update

At present, additional anecdotal evidence still suggests there is an association between J2b-Delta (x J2b-Zita) and the British Isles. Anecdotal evidence suggests that J2b-Delta might also be associated with Spain (but this needs to be confirmed). The project has further members that belong within J2b-Zita, and these members have diverse origins. There is no correlation between J2b-Zita and geographical origins within the British Isles.
The project currently suspects that J2b-Zita is not a phylogenetic sub-group of J2b-Delta. However, we do not yet have sufficient concrete evidence to reject the hypothesis. Rejecting the hypothesis that J2b-Zita is a phylogenetic sub-group of J2b-Delta purely on the basis of the geographical origins of the two groups would be unscientific.

We still need concrete evidence to distinguish J2b-Delta as a true phylogenetic grouping.
We also need concrete evidence to indicate whether J2b-Zita is a true phylogenetic grouping, and also whether it is within J2b-Delta, or phylogenetically separate from J2b-Delta. The results of SNP testing might in future give us the required concrete evidence to either confirm or reject these hypotheses.

If J2b-Zita is a phylogenetic grouping within J2b-Delta, then the diverse origins of the members within J2b-Zita would either indicate that the hypothesis that there is a correlation between J2b-Delta and the British is false, or it might suggest that J2b-Zita represents a back migration into continental Europe from the British Isles. At present, we'd consider the former more likely than the latter. Pursuing the latter without sufficient concrete evidence could be considered "scientifically suspect" (one feature of pseudo-science is dismissing all evidence that is contrary to the hypothesis).

If we assume that they are separate phylogenetic groupings, and that the correlation with the British Isles is real, then as already discussed, the interesting question to answer will be, - in which of the many historical migrations to the British Isles did J2-Delta men (or the progenitor of J2b-Delta) arrive?

In genetic genealogy circles, the most commonly touted hypothesis for the presence of J2 in the British Isles is from Roman soldiers. To examine this hypothesis we will compare the relative frequencies of clades in the British Isles with those in Italy*.

To examine this hypothesis, we will also need to evaluate on a finer scale, the geographical origins within the British Isles of J2b-Delta compared to the other clusters within J2. At present, the breakdown of region/sub-clade within the British Isles (derived from existing project members) is at left -->

In this diagram, for the haplotypes that are non SNP tested, the likely clade has been predicted. The percentages in white are for haplotypes that are most likely to be J2a, unlikely to be in J2a-Beta, but the M67 status is indeterminate. The pie that is situated between Ireland and the rest of the British Isles is for haplotypes for which a region within the United Kingdom/British Isles/ Great Britain is not specified.

It can be seen that there are greater proportions of J2b in England, and the proportion of J2a M67+ is higher in Scotland and Ireland. 50% of J2b haplotypes in the British Isles have a DYS 385a value of 13 or less (not all of these haplotypes are classified into J2b-Delta, as these figures include haplotypes with less than 37 markers also). There are 34 haplotypes that are likely to be J2b in the project that have origins that are stated to be a region outside of the British Isles. Approximately 33% of these have a DYS 385a value of 13 or less.

At left are the proportions found in Italy. Much of the J2* found in Italy falls within J2a-Beta. If we compare the proportions found in Italy with the proportions found in the British isles, the differences in the relative proportions might suggest that much of the J2b found in the British Isles came from a source other than Italy. (which might negate the Roman soldier hypothesis, or it might merely negate a hypothesis of Roman soldiers of Roman ethnicity).

On the other hand, it is important to consider that the haplotypes within the project for either the British Isles or Italy (or both) might not be representative of the true proportions that exist in those regions. The disparity seen between the British Isles and Italy might merely be a spurious result derived from low sample size and sampling bias.
Alternatively, it might be false logic to assume that the sub-clade ratio of J2 Y-DNA introduced by ethnically Roman soldiers would approximate the modern day proportions of sub-clades in Italy.

Four factors might cause a disparity:
1) The "Genghis Khan effect", ie. elite dominance, causing a greater percentage of descendants of one particular man or family line.
2) A bottleneck/genetic drift effect caused by a small pool/subset of individuals introducing Roman J2 Y-DNA, with the sub-clade proportions changing over time due to random genetic drift effects (eg. daughtering out), or in combination with an elite dominance effect;
3) The original pool of J2 Y-DNA introduced to the British Isles by Roman soldiers might not have been representative of males of Roman Ethnicity due to chance, and/or genetic heterogeneity
4) The proportions of J2 sub-clades have altered in Italy over the last several thousand years, so that the proportion of sub-clades in Modern Italian males is not representative of the proportions that were present in Ethnically Roman males 1600-2000 years ago.

We may have a better idea in the future, with the continued growth of the project. We should also further consider whether J2b-Delta is indeed also associated with Spain, and if so whether the progenitor of J2b-Delta might have been from Spain or from the British Isles (ie. did men with the J2b-Delta lineage migrate from Spain to the British Isles, or vice versa).

We would hypothesise that it is unlikely that haplogroup J2 was introduced to the British Isles solely by Roman soldiers. It was most likely introduced to the British isles on several different occasions, from different population movements. For any one haplotype, it would be difficult (if not impossible) to conclusively identify which population movement resulted in its paternal line ancestors arrival in the British Isles .

February 2008 Update
      (prepared in November/December 2007)

For this update, we will be again examining J2b-Delta (along with the rest of J2b), and we will also be examining J2a-Beta. We also add a few comments about J2a haplotypes that have the DYS 413 ancestral condition.

 • J2b

 • J2a-Beta (& J2a-Gamma)

 • J2a DYS 413 >18

In the sections below we show distribution maps of several clusters within J2b and J2a-Beta. Bear in mind that the areas of highest frequency do not necessarily indicate the point of origin of that cluster. Higher haplotype/subclade diversity is likely to be more indicative of origin.

J2b
(November/December 2007)

Below can be seen a version of the member map, that shows the the breakdown of J2b clusters. Keep in mind that the below constitutes a very small sample size, and may not be representative to the proportions of each cluster that might actually exist in the regions depicted below. In the map below, the faded red/pink portions represent the proportions of J2b haplotypes that are currently unable to be assigned to a cluster (12 marker haplotypes cannot be assigned to a cluster).

 

We currently suspect that most (if not all) in J2b-Alpha, J2a-Delta, and J2a-Zita are M241+. We no-longer think that M241 might distinguish either J2b-Delta or J2b-Zita from J2b-Alpha. We currently hypothesize that J2a-Gamma might represent haplotypes that are negative for both M241 and M205. The project currently has 3 haplotypes classified as J2b-Gamma - two from Germany and one for Yemen. We find the apparent diversity of J M12+ haplotypes in Germany quite interesting (perhaps this suggest that J M12 has been in that geographical region for quite some time).
More testing of the M241 SNP marker would assist in confirming this. Over the past 6 or more months we have communicated with a specific DNA company asking whether they can make M241 testing available. Hopefully there will be a greater availability of testing for M241 from that testing company after the new YCC tree is published.

J2b-Beta

 

J2b M205+ (represented by J2b-Beta cluster) is a clade for which very little is known. Virtually all of our knowledge of this clade is derived from what we have learned within this project. The map depicts what distributions are known (not shown is one haplotype from Mexico, and one haplotype from Russia which is suspected to be M205+). The greatest number in the project are from Greece, and the most divergent haplotype is from Serbia. This might suggest an origin from the Balkans (which perhaps isn't particularly illuminating, since the Balkans seems to be the region from which J2b as a whole is most frequent). It is perhaps relevant to again note the fact that Pericic et al. 2005 found approximately 50% of J2b in the Balkans to be M241-, with the greatest frequency of J2b M241- to be in Serbia. Unfortunately we cannot be sure what proportion of those J2b M241- might have been J2b M205+, since M205 was not tested. We can only speculate that perhaps a large proportion of those tested as M241- were M205+.
J2b-Beta cannot be conclusively identified at 12 markers (although haplotypes do have a tendency of having higher than normal DYS 385 values), and is identifiable by its values at DYS 454 and DYS 437. At the time of the February 2007 cluster analysis there were insufficient numbers of J2b-Beta with 37 markers to produce modals for the 26-37 marker panel. Modals for this panel have now been added (none of the markers in this panel appear to be significantly diagnostic). At present only one J2b-Beta haplotype has been tested at 67 markers.

 

J2b-Alpha

 

We can see here that J2b-Alpha has a relatively wide distribution. We also note that J2a-Alpha is the most frequent J2b-cluster in Germany. If we were to speculate about the arrival of J2b-Alpha in the British isles, we might be tempted to suggest either Roman troops of Germanic origin, and/or Anglo-Saxons. However, attributing specific haplogroup sub-clades and/or clusters with past nations and/or tribal groups is at best tenuous, and at worst potentially misleading. Hypotheses can often be mistaken as fact, which potentially leads to misinformation. The region with the highest frequency of a clade or cluster is not necessarily the source population.

J2b-Delta

A new observation that can be made with regard to J2b-Delta, is that all in J2b-Delta (& J2b-Zita) who have tested DYS 463, have values of 20. All J2b outside of J2b-Delta (& J2b-Zita) have a value of 22. As only 9 individuals within J2b have been tested for this marker (6 of which are J2b-Delta), its not certain whether this observation is significant, or whether an artifact of the small sample size. We thank all members who have tested for DYS 463.

A look at Y-search data suggests that the above observation may have been in part due to small sample size. Values of both 20 and 22 can be found in J2b-Alpha and J2b-Zita. However, virtually all in J2b-Delta do seem to have a DYS 463 value of 20. This might suggest that J2a-Delta and J2a-Zita are phylogenetically separate, and their clustering together is artificial (but on the other hand, DYS 463 is merely a STR - only new SNP markers will give us definitive answers). Again, the sample size is very low - and it is entirely possible that additional results might disprove the above observation. This all part of natural process of science.

 

It can be seen here that, by far J2b-Delta is most frequent in the British Isles, and England in particular. It can also be seen that J2b-Delta haplotypes have been found elsewhere, - Spain, Italy, Greece and Finland. Could it perhaps be that the higher numbers in the British Isles are merely the result of higher sampling from the British Isles? If the sampling elsewhere in the world was proportional to that of the British Isles, would we find the same frequencies of J2b-Delta elsewhere in the world?

If we look at how many people have tested with ancestral origins from the British isles (approx 33258), compared to how many have tested with ancestral origins from Finland (622), Greece (404) Italy (1660) and Spain (1642) - we could conclude that the frequency of J2b-Delta is lower in the British Isles compared to the rest of Europe, or (at best) is the same as the rest of Europe . A chi square analysis was insignificant (which means that the frequency of J2b-Delta is neither higher or lower in the British Isles, compared to elsewhere in Europe).
We should however remember that the sample sizes are very very small - too small to make any scientifically robust conclusions. Obviously, only a sub-set of individuals who have tested as J2 are included within this project, and we don't know how representative the current members are of all who have had their DNA tested, (nor how representative those tested as J2 are representative of J2 globally). Many more J2b-Delta haplotypes are required to make the comparison sufficiently robust.

Perhaps the frequency of J2b-Delta in the British Isles is not remarkable in itself, perhaps what we should be focusing on is the other clusters found elsewhere in the world that are at much lower frequencies in the British Isles. On the other hand, - even if the frequency of J2b-Delta is more-or-less the same throughout Europe - perhaps that in itself is remarkable.

If we look at the variation within J2b-Delta, we can calculate a cluster age of between 2000-4000 years (depending on the mutation rate used). We could try to relate J2b-Delta to some specific ancient population movement (for instance the Celts, Indo-Europeans, Anglo-Saxons, Roman troops). Any hypothesis devised would also have to account for the apparent absence of J2b-Delta in Scotland. At best - attributing specific haplogroup sub-clades and/or clusters with past nations and/or tribal groups is tenuous, and at worst potentially misleading. Hypotheses can often be mistaken as fact, which inadvertently leads to misinformation.

J2b-Zita

As mentioned in the October 2007 update, one member within J2b-Zita has DYS 413 values of 18 23. within the next couple of months we should have a better idea whether this was an isolated occurrence or whether it might be characteristic of a subset of J2b-Zita. At the moment there are insufficient numbers of J2b-Zita in the project to make any conclusions about its geographical distribution.

Overall, it can be seen that the results are suggestive of J2b reaching the British isles on more than one occasion. Any theories about the arrival of this clade would have to account for the different distribution patterns seen for each cluster (we should also remember that the markers used in most scientific studies do not give sufficient resolution to distinguish between the clusters). . Another observation that is clear is that the sample sizes are quite low. A far greater sample size with extended haplotypes is needed to determine whether we are starting to see real patterns here, - or whether the apparent patterns are just "noise".

J2a-Beta
(November/December 2007)

For J2a-Beta, we first examine in depth the clustering within J2a Beta, especially in relation to J2a-Gamma. We will then show the geographical origins of members within this cluster, and two of its subclusters..

In the October 2007 results analysis update, we stated that we had found that there was a cluster that contained both J2a-Gamma cluster haplotypes and J2a-Beta haplotypes in the 37 network diagram analysis. This created concerns about whether J2a-Gamma cluster (containing M67+ haplotypes) can be objectively distinguished from some haplotypes that are classified as J2a-Beta cluster (which conversely contains M67- Haplotypes). We promised that we would examine this finding in more depth at a supplementary analysis update. To do this, we first reanalysed the 37 marker haplotypes with splits tree. The results of the analysis with splits tree can be seen below (J2a-Beta/Gamma portion only shown).

Here there is better resolution of information than there was with the network diagram. Some of the J2a-Beta haplotypes that previously clustered with the J2a-Gamma cluster are now depicted in a subcluster with other J2b-Beta haplotypes. There are two haplotypes within the J2a-Gamma cluster that are known to be M67- and have previously been classified as J2a-Beta. Both these haplotypes are considered "outliers" within J2a-Beta. There are two haplotypes within the J2a-Gamma portion of the tree that have not been SNP tested, and at this stage we cannot be sure whether they are M67+ or M67-. We therefore cannot classify them as being in either J2a-Gamma or J2a-Beta (neither has been tested to 67 markers). In the past we had classified these haplotypes as being in J2a-beta, but we now suspect that they might be M67+. As a compromise, we are now classifying them as J2a-Beta/J2a-Gamma.

One conclusion on that can be readily made from this diagram is that - if we had done the February 2007 cluster analysis with all the haplotypes that are currently in the project, (November 2007) we probably would not have created Gamma cluster. The current Gamma cluster would probably have been created as a sub-cluster of J2a-Beta. We realise that our creation of J2a-Gamma cluster as separate from J2a-alpha and J2a-Beta was shaped partially by prior knowledge that haplotypes in J2a-Gamma are M67+, and haplotypes in J2a-Beta seem to be M67-.

The question now is - should we create the clusters totally "blind" to subclade status, reassigning those J2a M67+ haplotypes that are in J2a-Gamma cluster to J2a-Beta cluster, - or should we instead try to maintain the phylogenetic separateness of J2a-Gamma cluster from the M67- J2a-Beta haplotypes? We have decided to go with the latter option for now, - part of the reason for this is the fact that although J2a-Gamma haplotypes can be very difficult to clearly separate from some J2a-Beta haplotypes at the 37 marker level, - at the 67 marker level they are very clearly, and easily separated. See the colour coded version of the October 2007 67 marker network diagram below.

This emphasizes how valuable some of the markers in the 38-67 marker panels are in distinguishing different clusters. Haplotypes clusters that are difficult to separate at 37 markers are very easily distinguished at 67 markers.

Hopefully at the next analysis update we will be able to revise the cluster analysis. At present (by comparing 37 and 67 marker analyses using splits tree with the original 37 marker network diagram) it seems we might need to revise/reconceptualise J2a-Alpha and J2a-Delta. In terms of the former, - should our definition of J2a-Alpha include all J2a haplotypes that have the derived character state at DYS 413? (which would effectively mean that J2a-Beta is nested within J2a-Alpha). Or alternatively, should we let the 67 marker diagram define J2a-Alpha? (in which case, J2a-Alpha would most likely be the branch on the right of the 67 marker diagram above). Alternatively we could do both - redefine J2a-Alpha as J2 DYS 413 derived, and redefine J2a-Delta as the right branch on the 67 marker diagram.

The very fact that haplotypes from different subclades can cluster together emphasizes the fact that haplotype clusters are not equivalent to phylogenetic clades. The only way that we can be sure that a cluster really does represent a true biological grouping is through the discovery of SNP markers that correlate to the clusters.
We outline this more on the new "project hypotheses" page.

We should also recap on the fact that that DYS 450 is of high diagnostic value for J2a-Beta. Outside of J2a-Beta (inclusive of J2a-Lambda and J2a-Theta), the usual value is 8 and within J2a-Beta the usual value is 9. This pattern holds in most instances, - but (as always) there are exceptions. We know of one haplotype with a DYS 450 value of 9 that is evidentially not in J2a-Beta (it is in fact in J2a-Zita and at 67 markers clusters with another J2a-Zita haplotype, well away from J2a-Beta).

Geographical Distribution of J2a-Beta
 

Above shows the proportion of haplotypes for each country/region that are in J2a-Beta (including J2a-Theta and J2a-Lambda). Below are seperate maps for J2a-Theta, J2a-Lambda and J2a-Beta (non inclusive of those in J2a-Theta, J2a-Lambda).

J2a-Beta (non inclusive of J2a-Lambda and J2a-Theta)

 

When looking at the proportions above you should remember that J2a-Beta is in a sense "paraphyletic". ie. J2a-Beta is likely to contain many sub-clusters and the graph above excludes two of the more important sub-clusters. The map above shows the highest proportion in France, but given the small sample size, it's uncertain whether this is of any significance.

 

 

It can be seen in the map above that the highest frequencies of J2a-Lambda appear to be in Germany and in Italy. What we can't tell for sure is whether it is more likely to have originated in the Alpine region between the two regions, in the region of Italy, or the region of Germany (or perhaps a different geographical region entirely). Within the British Isles, the region with the highest proportion of J2a-Lambda is Scotland. Of note is the fact that in the 37 marker tree, a sub-cluster of J2a-Lambda seems to be found almost exclusively from the region of Germany/Netherlands. We again caution, that because of the relative small sample size, we are not sure whether these observations are significant, or whether they are an artifact of the small sample size.

J2a-Theta

 

J2a DYS 413 >18

J2a (DYS 413 >18) has been referred to as a cluster (outside of this project). That characterisation is quite misleading, and in fact, DYS 413 >18 haplotypes are quite diverse, and do not fall within a single cluster. In a Phylogenetic sense we are likely to find that J2a (DYS 413 >18) is paraphyletic. In other words - Some of these clusters may be more related to haplotypes that are DYS 413 <18, than they are to other clusters of DYS 413 >18 haplotypes. There are at least three distinct groups (one of which is J2a-ε), and the non J2a-ε haplotypes are predominantly from either Germany or France.

It will be interesting to discover whether there are any J2a (DYS 413 >18) haplotypes that are positive for S57 (see this entry on the updates page).

The distribution of J2a DYS 413>18 is of theoretical interest, because we would expect higher haplotype diversity in the areas where a particular clade originated. Earlier scientific studies have hypothesized that J2 originated in the Middle East, north of the Zagros mountains. If this is true, then we would expect a greater diversity of J2 DYS 413>18 haplotypes in the Middle East. At present it is difficult to interpret the distribution found - but its important to again remember that the geographical origins of people that have had their DNA tested are not representative of the global population, and biased towards Western Europe. If testing was more proportional, then we would (in theory) expect more J2a DYS 413 >18 haplotypes in the Middle East/Eastern Europe. It would be interesting in the future to find out whether this is indeed true.

In future we will add information for additional clusters.

Review of the Cruciani et al. 2007 paper
       (added April 2007)

In summary, Cruciani et al. (2007) suggest that both E3b-V13 and J2b were involved in a demographic expansion from the Balkans at the beginning of the Balkan Bronze age.
This hypothesis is consistent with the deduction made by the project at the time of its February 2007 cluster analysis, that we cannot assume that J2b was in any way involved in the demographic expansion associated with Neolithic farmers.

We would further hypothesize that the population histories of J2a and J2b are quite different and separate Our specific hypotheses are listed below:

J2a was prominent in the demographic expansion associated with the introduction of agriculture from Anatolia.
J2b was not prominent in this demographic expansion, and was involved in a later demographic expansion in Europe, which may or may not have been the same demographic expansion that is responsible for the expansion of E3b-V13. It may also have been involved in an earlier demographic expansion that introduced J2b into both Europe and India from an as yet unidentified source population.

There are several academic research papers that have recently published estimated coalescence ages for J2b. Both Cruciani et al. 2007 and Peričić et al. 2005 have published papers that date the demographic expansion of J2b within Europe as being approximately 3000 - 4500 years ago. On the other hand, the Sengupta et al. 2006 paper estimates the age of J2b in both Dravidian and Indo-European Castes as 12000 years (± 3400), and in Austro-Asiatic Tribes as 3600 years (± 1700).
As a comparison, Sengupta et al. also estimate the age of R1a as 12000 years, and Peričić estimates the age of R1a in Europe as 15800 years ± 2100.

It is important to recognise these dates as representing the major demographic expansions of the sub-clade, rather than representing the age of the sub-clade per se. Even if the common ancestor for most of J2b in Europe lived 4500 years ago, it's important to remember that the divergence between J2a and J2b must have occurred earlier than the overall coalescence ages for the lineages within J2a. Therefore the overall progenitor for J2b must have lived well before the coalescence ages for J2b within Europe. Interestingly, the projects estimate of the age of J2b M205 from project members genetic distances estimates a coalescence just within the overall coalescence estimate of 4500 years for J2b in Europe.

It's also important to remember that the estimated age of a haplogroup in a specific geographical region does not mean that the clade has been present in that geographical region for all of that length of time. As a comparison, - if the R1b variation in the USA (or Canada, or Australia, or New Zealand etc.) was calculated, the calculated coalescence is likely to be far greater than 500 years. Some of the variation seen is derived from the variation that may have been present in the original source population.

All three papers use the methods of Zhivotovsky et al. (2004) as the basis of their calculation methods. Both Sengupta et al. 2006 and Cruciani et al. 2007 used a slightly modified version of the method, to take into account the disparities seen by mutation rates inferred from population data, and mutation rates inferred from father-son studies.
In Zhivotovsky et al. 2006, it is stated that the disparity between mutation rates inferred from population data, and mutation rates derived from father-son studies, is due to demographic factors. Variation within a clade is reduced by genetic bottlenecks, that may result from demographic factors. These demographic factors will differ between different populations, meaning there is no constant correction factor for the disparity seen between father-son mutation rates and those inferred from population studies.
The relative similarity between the coalescence dates for R1a between the Sengupta et al. 2006 and Peričić et al. 2005 papers suggests that the calculation methods they use are more-or-less comparative, and therefore the difference in coalescence dates for J2b is due to different expansion times and/or demographic events that have impacted on haplotype variation.

Paradox - greater microsatellite variation in Indian J2b, but greater phylogenetic variation in European J2b.

One question that could be asked is - Is the J2b in India derived from the J2b in Europe, or vice versa? We would hypothesize that it is derived from neither, and both are derived from an as yet unidentified source population. We can also hypothesize that after moving from the original source population, both separate populations of J2b have experienced different demographic factors which have resulted in different impacts on the microsatellite and phylogenetic variation. This hypothesis is the best explanation for the apparent paradox of greater microsatellite variation in Indian J2b, yet greater phylogenetic variation in European J2b. Our rationale for this is as below:

Assuming that the coalescence dates determined by all researchers have been calculated in a consistent unbiased manner between haplogroups, we could easily conclude that J2b in Europe is derived from the J2b in India. On the other hand, according to Sengupta et al. 2006, all of the J2b in India is M241+, whereas Europe has both J2b M241+ and J2b M241- (as well as J2b M205+). We would therefore expect greater variation in Europe (since there is greater phylogenetic diversity).
We can also conclude that the J2b in India is not derived from the J2b in Europe. This can be deduced from the greater coalescence date for J2b in India, and also the fact that E3b is either absent or negligible in populations within India. Therefore, the J2b in India is from a source population that pre-dates the Bronze age expansion of J2b and E3b in the Balkans.
The disparity may be due to different demographic events leading to reduced variation in Europe (eg. loss of a greater percentage of lineages due to more daughtering out, bottlenecks etc).
In simple terms - Perhaps many of the branches of J2b in Europe were pruned out. Therefore we hypothesise that the J2b in both Europe and India are derived from an as yet unidentified common source, with more of the original J2b M241+ microsatellite (ie. STR) variation preserved in India compared to Europe, due to differing demographic effects.

Another factor to consider is that there are distinct differences in the DYS 461 values between European M241+ and Indian M241+. Most European M241+ haplotypes have DYS 461 values of 10, and most Indian M241+ haplotypes have DYS 461 values of 9. We cannot as yet determine which marker value is likely to have been the ancestral marker value. It would be interesting to know what marker values M205+ haplotypes have (as yet, none of the projects M205+ haplotypes have been tested for DYS 461).

We should also consider the fact that it may be false logic to conclude that the calculated coalescence age for J2b in India is accurate even if the European and Indian coalescence ages for R1a are more-or-less equivalent. Zhivotovsky et al. 2006 states that the demographic factors that cause disparities between the mutation rate inferred from population data, and the mutation rate inferred from father-son studies will differ between different haplogroups, and different populations. The calculation method used by Zhivotovsky et al. (2004) was derived from studies examining microsatellite variation in populations with known foundation times (Bulgarian Gypsies, and New Zealand Māori). There will be some populations that experienced greater bottlenecks than these populations (thus the Zhivotovsky et al. (2004) method will underestimate coalescence age), and there will be populations that experienced fewer bottlenecks than these populations (thus the Zhivotovsky et al. (2004) method will over estimate coalescence age).

The project is currently working on formulating further hypotheses regarding the demographic history of J2b. We are however aware of the fact that it is very difficult to conclusively attribute any haplogroup to any specific historical groups/ population movements. We will therefore refrain from hypothesizing that any specific historical/archeological groups were responsible for the demographic expansion of J2b (regardless of how tempting it might be to attribute J2b to groups such as Indo-Europeans).

 

 

*Most Roman soldiers that were in Britain were unlikely to be ethnically Roman. Soldiers in the Roman Empire were recruited from many different regions. Because these soldiers were from a variety of different ethnic backgrounds (each with differing haplogroup proportions), it would be very difficult to objectively test the hypothesis of Roman soldiers of diverse ethnic origin. This is because we'd effectively need to compare the sub-clade proportions in Britain with the sub-clade proportions in much of Europe where Roman soldiers were recruited. This means we would not be able to easily differentiate this comparison, from an alternative hypothesis of J2 being introduced to Britain multiple times, from multiple geographical areas.
On the other hand, it is somewhat more possible to objectively test the hypothesis of ethnically Roman soldiers (even though it is from the outset unlikely that this hypothesis would be supported).
Our comparison tentatively rejects the hypothesis of the J2 in Britain resulting from ethnically Roman soldiers, but does not negate the hypothesis of Roman soldiers of differing ethnicities. We do however suspect that the J2 in Britain did arrive at different times (but of course cannot yet conclude it as fact). We will in future outline suitable hypotheses outlining how we can separate this hypothesis from the alternative hypothesis of Roman soldiers of differing ethnicities.

Creative Commons License
This work is licensed under a
Creative Commons Attribution-No Derivative Works 3.0 License.
This work can be freely cited, if it is attributed to:
The J2 Y-DNA project
http://www.j2-ydnaproject.net

Angela Cone - Co-administrator from mid 2006 - mid 2008
Administrator from mid 2008 - Present
Click here to read about Angela.

Costa Tsirigakis - Founder J2 Y DNA project & admin from 2006 - mid 2008